Do Tough Prosecutors Save Lives?
Concrete Evidence (March 2, 2026)
Back in October, this column took a look at “progressive prosecutors.” These folks tend to get elected in blue cities with promises of decarceration, but new research had shown them to have little impact on imprisonment—or, thankfully, crime. Their gentler approach seemed largely confined to nonviolent misdemeanor cases.
A new working paper, already spotlighted over at Marginal Revolution, takes another look at the difference prosecutors makes. But instead of measuring the impact of trendy new “reform” prosecutors, it looks at the more mundane situation of close elections between Republicans and Democrats.
This time, the harsher candidates are the ones who improve things. Electing a Republican prosecutor reduces mortality almost 7 percent among men aged 20 to 29, primarily through reduced gun deaths. Let’s dig in a bit.
The study’s design is key. Rather than comparing deep-red areas with deep-blue ones—an approach that would inevitably be muddied by differences in culture, demographics, and other confounding factors—the authors focus on close elections. Places where a Republican prosecutor wins 51-49 shouldn’t look dramatically different from one where a Democrat wins 51-49. If death rates diverge right at the 50 percent threshold, the election result itself is a plausible explanation. The authors’ model thus checks for a “discontinuity” in death rates at that 50 percent mark and finds the aforementioned 7 percent mortality reduction for young men in areas where the Republican candidate narrowly prevailed.
Breaking down the change, the authors largely find a mix of reductions in gun homicides for black men and gun suicides and accidents for white men, with little impact on non-gun homicides. Of course, that’s a little odd: It’s easy to see how prosecutors can reduce crimes like homicide, but why do they seem to have such an impact on gun deaths across the board, including non-crimes?
The authors explore this question using some other data sources. Republican prosecutors appear to increase prison admissions for black men, incapacitation from which can explain part but not all of the mortality decrease for this demographic.
By contrast, these prosecutors increase criminal convictions for both racial groups, and across a wide range of different offense types. The authors’ theory: The convictions, more so than the incarceration, might be driving the mortality effect.
Specifically, felony convictions (and certain misdemeanors) come with a prohibition on owning guns. Such a prohibition will cause someone to fail a background check if he tries to buy a weapon from a licensed dealer—or even from a private seller, in states that require checks on all sales—and he’ll risk serious punishment if he buys a gun on the black market.
In an admittedly back-of-the-envelope calculation, the authors estimate that one gun death is prevented “per 82 restrictions among Black individuals and per 91 restrictions among White individuals.”
Is gun control for convicts really that effective—not only homing in on an extremely high-risk population, but successfully deterring them from keeping firearms around, such that more than one life is saved for every 100 people banned from guns this way?
I’m a bit dubious, given America’s robust black market in firearms and the tendency of criminals not to follow laws. But if so, expanding the list of offenses that gun prohibitions apply to, as well as consistently pursuing these convictions where possible, sounds like a good idea—despite inevitable pushback from both the left and the right.
I hope this working paper inspires some more research. Critics and peer reviewers should kick the tires hard before it’s published. And it’s worth reconciling this result with conflicting ones in the literature—fleshing out when prosecutors save lives by being harsh, and when they can afford to be lenient.
From the Manhattan Institute
Selective high schools in Chicago “produce better results, at a lower cost, than the system as a whole,” finds Renu Mukherjee.
Other Works of Note
Keep your wild eyes on the road: A new working paper finds that when major new pop releases come out, traffic fatalities rise. The authors treat this as a sign of people futzing with their phones while driving, but maybe everyone is just rocking out too hard to the new T-Swift behind the wheel?
If a company’s costs increase from $50 to $55, and the firm then “passes through” the increase to consumers, does that mean the final price goes up by $5—or by 10 percent, maintaining a consistent ratio of final price to cost? Apparently, studies of “pass-through” tend to assume the latter, and if you assume the former instead, pass-through rates are much higher than previously thought.
Speaking of pass-through, do metal prices affect not only car prices, but the terms of auto loans?
Manhattan Institute’s Judge Glock on why clean water is so expensive.
Public opinion data on legalizing psychedelics.
People are less likely to move from County A to County B—or to live in one and work in the other—if the two counties are in different states, even apart from the actual distance between the two.
E-cigarette taxes seem to reduce use even among youth who aren’t supposed to be able to buy the products directly.
How Obamacare affected uninsurance among folks who lost their jobs.
A randomized experiment gave families of preschoolers access to a math app and “analog” math materials. “On average, neither intervention significantly increased children’s math skills.”
Analyzing AI use (and transparency about it) in academic journals. And could AI exacerbate cognitive inequality?
Law-review article on data scanning and the Fourth Amendment.
From Justin Nix, an RSS feed automatically tracking research on police shootings.
Both pro-gun and gun-control groups increase political donations in districts that suffer school shootings.
Nerd alert: The Congressional Budget Office dumped a bunch of its modeling code onto GitHub.
The occupational integration of black immigrants.
Interesting idea for addressing nonresponse bias in surveys.
When a Danish reform “earmarked” paternity leave for dads rather than letting couples allocate it freely, the amount of leave taken shifted from moms to dads as expected, but parents’ satisfaction with leave declined. And Spain’s “baby bonus” had “little impact on children’s health and test scores.”
“Is Physical Unattractiveness a Risk Factor for Sexual Violence Perpetration? Evidence from the U.S.” Who said academics won’t study taboo subjects anymore?
Steve Sailer often makes the point that depolicing, or at least taking away police ability to look for illegal guns also takes away disincentives for knuckleheads to carry guns. As would lenient judges.
Expanding list of offences which prohibit gun ownership might have a decreasing marginal reward. At the moment the cut-off are individuals with moderate probability of harm or self-harm. If the cut-off is moved to low probability, the marginal effect will be difficult to observe.